1. Field of the Invention
The invention relates generally to arc fault detectors and, more particularly, is a device for sensing the temperature of insulation to detect when it is about to fail, and, therefore, for predicting and preventing a fire based on the resultant electrical arcing.
2. Description of the Related Art
U.S. Pat. No. 5,157,380 to Braun et al. discloses the use of a metal oxide semiconductor (MOS) detector to analyze the exhaust of a turbo generator for overheated insulation. This device detects carbon monoxide and methane breakdown products for the insulation and requires on-line reference gas, automatic valves, pumps and associated control electronics. Stability of the technique is a problem which is why on-line calibration is necessary. MOS detectors sense many other gases in addition to overheated insulation products. In short, they are not very specific when they are exposed to the random contamination experienced on a ship. They will alarm for jet engine exhaust, diesel exhaust, paint fumes, freon, and many other gases. This makes them most suitable in locations where the atmosphere is clean and closely controlled. On board a ship they give too many false alarms to be useful. The complexity and cost of Braun et al.'s technique make it suitable only to protect centralized high dollar value items such as very large motors and generators.
U.S. Pat. No. 5,362,568 to Dietz et al. requires treating all surfaces to be protected with tagged compounds. If different materials in the turbo generator are treated with different tracers one can determine which material is overheating before you open the generator for repair. This is an advantage over the bulk detection method by Braun et al. Dietz et al. then perform standard analysis using a standard gas chromatograph. This requires the used of two bottles of high pressure gas, one to calibrate (called "reference gas") and the second to operate the analyzer. Additional pumps, valves, and control electronics are also required. This is a typical batch analysis and is not continuous. Because of the high complexity and cost of this system it is warranted only when protecting centralized high dollar value items such as the large motors and generators. This method will not work with existing equipment since the original insulation would not contain the needed tracers.
U.S. Pat. Nos. 3,916,671 and 3,807,218, to Carson et al. are for a gas sampling method and gas chromatograph analysis of the cooling gas from the generator to identify pyrolysis products which would identify the overheated component without disassembly of the machine. All of the comments above made about the patent to Dietz et al. apply. This is a large expensive off-line analyzer suitable to expensive generators.
U.S. Pat. No. 4,117,713 to Phillips et al. refers to a particulograph. The drawings attached show the need for several valves, a detector of some sort and an analyzer gas supply. The discussion on size, complexity, and cost for the gas chromatographic methods applies here.
U.S. Pat. No. 4,101,277 to Hickam analyzes the ratio of oxygen to nitrogen rather than the direct overheating gaseous products. Hickam states that even in the hydrogen cooled generators there are traces of air. Since the ratio of oxygen to nitrogen in air is stable and known it should be the same ratio in the generator. If the ratio changes then oxygen must have been consumed in the pyrolysis of insulation which was overheated. Technically this does not identify overheated components, it determines if the bulk heating exceeded the oxidation temperature of the materials. It is probably less expensive than the gas chromatograph methods, but will not furnish any information about which material is overheating.
U.S. Pat. No. 3,427,880 to Grobel et al. discusses the use of ionization chambers to detect the pyrolysis products of overheated insulation. The ionization chamber is of a different type than the one Applicants use. Grobel et al. uses Thorium 232 and coats the surfaces to be monitored.
U.S. Pat. No. 4,121,458 to Fort points out that the invention of the '880 patent to Grobel et al. has problems due to changes in gas pressure, gas purity, gas flow rate, and contamination of the radioactive source. Fort then describes the use of dual ionization chamber which addresses some of the problems.
U.S. Pat. No. 4,364,032 to Narato claims that increased sensitivity can be had by integrating the rate of pyrolysis product generated over time. This works only on sealed systems and only if the mass of protected insulation is known and accounted for in the calibration curve. If one knows that the box to be protected contains 5 lb. of insulation then one can analyze the total amount of insulation lost per time. If one presumes that the insulation loss is evenly distributed over the total amount of insulation present then a threshold can be set, say 1.0%, at which an alarm will be set. In machinery with forced cooling one can presume that the heat is evenly distributed by the cooling medium over the entire machine. In this manner one can justify alarming upon the total amount of insulation lost.
The method of Narato will not detect gross overheating of a single small spot. For instance, one can set the alarm at an effective total temperature of 5 watts per square foot times 10 feet or 50 watts total heat. A hot spot of 100 watts over a surface of 0.1 square foot would look like 10 watts total heat and not alarm. However 100 watts is considerably over the 50 watts selected for the alarm. Therefore in many cases the integration method creates problems during actual application. The integration method means that there are an infinite number of surface area times heat times time which will produce the same integral value. This means small very hot spots can be missed.
All of the above patents generally relate to detection of overheated insulation in large expensive generators. These units can cost several million dollars and frequently are larger than the typical office. Repairs are expensive and down time is lost revenue. If one can detect an imminent failure, one can perform repairs before the damage and cost are catastrophic. If one can identify which part is failing before opening the machine, one can have the parts on hand before bringing the machine off line. This reduces lost revenue due to down time. Hence the interest in coating parts with tracers or identifying the composition of the failing part from its off-gassing. If the possible lost revenue is thousands of dollars per hour, the size and cost of the analyzer to reduce down time is not of major importance.
The methods in the above patents are generally applicable to gas cooled equipment. The circulation of the above cooling gas is used to transport the pyrolysis byproducts to a point where they can be sampled. The detectors are not suitable for actual insertion into the equipment being monitored.
In summary, the above systems all have many similar disadvantages:
The items are not suitable for use outside a closed environment. They are not specific when they are exposed to random fluctuations, such as jet engine exhaust, diesel exhaust, paint fumes, freon, and many other gases. PA1 The systems depend on the background gases being known and constant to prevent false alarms. PA1 They cannot be installed into the equipment actually being monitored. They require circulation of cooling gas to transport the pyrolysis by-products to a point where they can be sampled. PA1 The units are only cost-efficient when protecting high value items. None of the items have a low cost per item protected. PA1 The items require modifications to the existing insulation being detected; the material must be inundated with tracers or the use of specific coatings. PA1 Most of the units involve consumables, such as the calibration gases. These gases must be replenished in order for the system to work properly. PA1 detection of the temperature of the insulation with a device which is small and can be installed inside existing enclosures that contain wires. PA1 detection of the temperature of the insulation in open enclosures without the use of forced circulation of cooling gas. PA1 detection of the temperature of insulation without modification of the existing insulation to be protected. PA1 detection of the temperature of the insulation in systems by one system that can be used in different environments; the device can be used in ships and aircraft, as well as on land-based systems. PA1 detection of the temperature without the use of consumables such as tracer gases in order to detect the state of the insulation. PA1 protection for items which do not have high specific costs associated with them due to the inexpensive nature of the invention. PA1 networking of the signals for multiple detectors to allow protection of many electrical enclosures. PA1 detection of particles in the invention's ionization chamber provides a much higher level of immunity to false signals than does the sensing of gases as is done in the prior art.